基于HGWO-SVM的燃煤电站锅炉受热面积灰预测

Prediction of the heating surface pollution in coal-fired power plant based on HGWO-SVM model

针对燃煤电站锅炉受热面以积灰监测模型指导吹灰操作效率低的问题,建立混合灰狼算法(HGWO)优化支持向量机(SVM)的预测模型,实现对省煤器吸热量的实时预测。首先选取与省煤器吸热量高相关性的输入变量,搭建SVM预测模型;其次利用HGWO算法对SVM的核函数参量寻优,并将最优参量赋给SVM模型进行训练,完成对省煤器吸热量的预测;最后将预测结果带入热力学公式计算得到清洁因子,以表征省煤器的积灰程度。以浙江某燃煤电站660 MW锅炉机组为例,将实测数据作为样本进行训练和验证。通过与传统模型预测结果进行对比,该模型训练时间更短,预测精度更高。

Aiming at such problems of the low efficiency of soot blowing operation guided by the pollution monitoring model of heating surface in coal-fired power station boiler,Building the prediction model of HGWO-SVM,real on-line prediction of heat absorption of economizer.Firstly,the input variables with high correlation with the heat absorption of economizer are selected to build SVM prediction model.Secondly,the hybrid GWO was used for parameters optimization of SVM,then the optimal parameters are assigned to SVM model for training to complete the prediction of heat absorption of economizer.Finally,the predicted results are introduced into the thermodynamic formula to calculate the cleaning factor,which can be used to characterize the ash deposition degree of the economizer.Taking a 660 MW boiler unit of a coal-fired power station in Zhejiang Province as an example,the measured data are taken as samples for training and verification.Compared with the prediction results of the traditional model,This prediction model has shorter training time and higher prediction accuracy.